PBM Bear Thesis Update: The AWP Alternatives Will Soon Be Here

Print Friendly, PDF & Email

Richard Evans / Scott Hinds / Ryan Baum


203.901.1631 /.1632 / .1627

https://twitter.com/images/resources/twitter-bird-blue-on-white.png revans@/shinds@/rbaum@ssrllc.com


May 28, 2013

PBM Bear Thesis Update: The AWP Alternatives Will Soon Be Here

  • The Center for Medicare and Medicaid Services’ (CMS) final rule replacing average wholesale price (AWP) as a basis for reimbursing pharmacies for generic drugs dispensed to Medicaid beneficiaries will be published this August. The rule will be effective as early as this September, but no later than December of 2014. Our best guess is that the rule takes effect in either January or June of 2014
  • To process Medicaid prescriptions, pharmacies nationwide will have to include the AWP alternatives (average manufacturer price or ‘AMP’; and/or national average drug acquisition cost or ‘NADAC’) in their information systems – thus making AMP and NADAC available for commercial (e.g. employer-sponsored) drug benefit contracts
  • AWP bears no consistent relationship to pharmacies’ costs of acquiring generics; two government studies cited herein show that AWP-based reimbursement tends to exceed true acquisition cost by roughly 400 pct. AMP and NADAC both are closely related to true acquisition costs; these same studies show AMP-based pharmacy reimbursement exceeds true acquisition cost by only about 35 pct
  • We’re convinced commercial plan sponsors will insist on AMP or NADAC as reimbursement benchmarks once these are available in pharmacy information systems, that this will reduce generic dispensing margins throughout the drug trades (PBMs, retail, wholesale), and that PBMs will be most negatively affected
  • A common counter-thesis is that PBMs will simply shift to the new benchmark but price at a level that maintains to total gross margin. We believe this is unlikely for several reasons, primary among these being the tendency of AMP or NADAC to eliminate much of what makes newly launched generics so particularly profitable
  • Because AWP is not tied to acquisition costs, as acquisition costs fall in the first months and quarters following the new generic’s launch, pharmacy margins rise (the AWP-based reimbursement remains constant, even while the cost of the generic to pharmacies is falling). AMP and NADAC are tied to acquisition costs, thus as the cost of a generic to pharmacies falls (as it tends to do dramatically early in the new generic’s life), the AMP or NADAC based pharmacy payment also falls, narrowing the opportunity for outsized generic margins in the months and quarters following launch of a new generic


In February of last year CMS[1] published a draft final rule[2] addressing, among other related items, the calculation of Federal Upper Limits (FULs) for Medicaid drugs purchased at retail. Earlier this year, CMS posted[3] a final action date of August 2013 for the proposed rule

The effective date of the rule is not specified[4]. The effective date can be as early as 30 days following publication of the final rule, meaning the new basis for calculating Medicaid FULs could be in effect as early as this September. In the draft rule, CMS proposes allowing the territories five calendar quarters to comply with provisions of the rule, on the basis that the territories will require more lead time than the states. Because of this, we view five calendar quarters after August 2013 (i.e. December 2014) as the extreme outer bound for the effective date of the rule. Our expectation is that the effective date will be either January 2014, or July 2014 (the start of the states’ fiscal 2015)

Key Provisions; Impact on Generic Dispensing Margins for Medicaid Rx’s

States’ reimbursement for generic drugs dispensed at retail to Medicaid beneficiaries must remain (in aggregate) at or below the FULs set by CMS. FULs currently are set at 150 pct of the lowest published compendium price for a given generic; the compendium price is generally published as ‘average wholesale price’, or ‘AWP’. For generics, AWP bears very little relationship to true underlying acquisition costs[5]; in recent studies both GAO[6] and OIG[7] found that AWP-based FULs result in pharmacy payments that exceed true acquisition costs by roughly 400 pct, on average

The rule proposes to calculate FULs using actual acquisition costs or ‘AAC’, and specifically proposes to use average manufacturer price (‘AMP’) as an estimate of AAC. AMP reflects the price at which manufacturers sell direct to retail pharmacies, or to wholesalers for distribution to retail pharmacies – and accordingly, AMP is (unlike AWP) closely related to true acquisition costs. For drugs to be eligible for Medicaid reimbursement, manufacturers are required to report AMP to CMS on a monthly basis within 30 days of the close of each month. The same GAO and OIG studies referenced above found that AMP-based FULs were on average more than 60 pct lower than AWP-based FULs, and exceeded true acquisition costs by roughly 35 pct

Optional v. mandatory

States have had the option of moving away from AWP-based FULs for more than a year; as of March 31, six states have officially done so (Alabama, Colorado, Idaho, Iowa, Louisiana, Oregon)[8]

Crucially, once the new rule is in effect, states will no longer be able to rely on AWP as a primary pricing benchmark. States must receive CMS approval for their Medicaid plans on a periodic basis, and CMS alludes in the rule that conversion to AMP-based FULs will be a requirement for state plan approval. Even if CMS did not administratively enforce the switch to AMP-based FULs, without making the switch states would be hard pressed to comply with the pre-existing requirement that their reimbursement for Medicaid retail prescriptions remain at or below CMS’ specified FULs in the aggregate. Because CMS clearly is switching to an AMP basis, and because AMP-based FULs are +/- 60 pct lower than AWP-based FULs, a state that remained on an AWP basis would have no way of knowing[9] whether its Medicaid reimbursements were above or below the CMS FUL maximum

The clear implication is that once the rule is final, states will have no choice but to switch to AMP-based reimbursement

Commercial Impact #1: AWP is replaced as a benchmark for commercial contracts

Commercial pharmacy benefit contracts tend to use AWP as a pricing benchmark, as do contracts among various members of the prescription supply chain (e.g. wholesalers and retailers). As we’ve long argued, AWP (bearing no predictable relationship to acquisition costs) creates an information asymmetry that favors the seller at the expense of the buyer, particularly when the buyer (e.g. an employer) is not a direct participant in the drug supply chain, and so has no direct knowledge of actual transaction costs

For the moment commercial plan sponsors cannot insist on alternatives to AWP, as (outside of the six ‘early-adopter’ AAC states) the alternatives are impossible to administer. AMP and NADAC[10] are viable alternatives to AWP that both closely reflect true acquisition costs, and these have both been published by CMS for more than a year. However neither AMP nor NADAC are consistently present in retail pharmacy information systems, thus despite being publicly available in the sense that they can be referred to via CMS’ website, neither is practically available in the immediate context of pricing / adjudicating a retail pharmacy claim

The states’ shift to AMP-based FULs for Medicaid eliminates this critical rate limit. Once CMS’ final rule is in effect, all states will rely on AMP-based FULs, and any pharmacy hoping to fill even a single Medicaid prescription will need the ability to price / adjudicate prescriptions using AMP – thus it’s a reasonably safe bet that standard information systems for retail pharmacy will be loaded with AMP (+/- NADAC) data. Once this occurs, AMP and NADAC become viable benchmarks for commercial contracts. Because AMP and NADAC closely track acquisition costs, and thus very nearly eliminate the information asymmetry that places plan sponsors at a disadvantage, we expect plan sponsors to insist on AMP and/or NADAC as alternatives to AWP as soon as AMP and/or NADAC are available

Commercial Impact #2: Pressure on generic dispensing margins

We believe the larger PBMs are very heavily reliant on generic dispensing margin[11]; in fact before its acquisition by ESRX, we showed that generic margin accounted for all, or very nearly all, of MHS gross profits[12]. This was inherently rational; because AWP enabled[13] outsized generic dispensing margins, it made sense for MHS to sacrifice other potential sources of gross margin (e.g. brand dispensing margin) in order to maximize its exposure to generics, particularly at mail. However once AWP is replaced, a key enabler of outsized generic dispensing margins is eliminated. Using AMP or NADAC, plan sponsors will for the first time be able to negotiate generic dispensing margins on a cost-plus basis, and if desirable or necessary this can even be done on a drug-by-drug basis

We believe this naturally leads to lower average generic dispensing margins; recall that in Medicaid, AWP-based FULs are 400 pct above true acquisition costs, and AMP-based FULs are more on the order of 35 pct above true acquisition costs. Commercial buyers may or may not apply the same multiples to AMP to calculate reimbursement (the proposed CMS multiplier is 1.75), though we don’t see any reason to expect commercial plan sponsors to agree to substantially higher dispensing margins than are paid by Medicaid

The case is often made to us that PBMs will simply replace AWP with AMP but maintain total gross margin by spreading margin more evenly across brands and generics. We believe this is incorrect for three reasons:

First, we believe PBM gross margins in the aggregate are ‘non-economic’; i.e. that the presence of AWP results in total gross margins that are higher than they otherwise would be. Elsewhere we’ve shown that the retail networks and mail dispensing services of a typical PBM can be replaced by a combination of existing closed loop payment processors and existing drug distribution centers, and that the costs of operating this alternative network are at least 16 pct below those charged by a typical PBM[14]

Second, we believe the PBM industry is under considerable pricing pressure, and that this is reflected by the facts that the two largest players are stealing share from one another by targeting large accounts, that two ambitious up-and-comers have achieved critical mass, and that this is all playing out against a backdrop of decelerating prescription volumes (Exhibit 1)

Third, even if plan sponsors somehow agreed to switch AMP for AWP at levels that preserved current generic dispensing margins, margins on new generics almost certainly would fall – and it’s abundantly clear that new generics account for a very large percentage of total generic margin. Under the current AWP-based benchmark, the contract price between plan sponsor and seller (PBM, retailer, etc.) remains constant for some period following launch of the generic, even though the seller’s cost of acquiring the generic is falling rapidly as more suppliers of that generic come on line. Thus under the AWP benchmark sellers see a period of expanding margin on new generics, and this continues until a point at which the AWP-based price is replaced by various versions of maximum allowable cost (‘MAC’). State Medicaid agencies will ‘MAC’ generics once acquisition costs have fallen and stabilized; similarly commercial payors (PBMs) eventually will ‘MAC’ generics (for a fee) on plan sponsors’ behalf[15]. We use the number of months between first generic launch and first appearance of a federal FUL as a very rough proxy for the number of months between first generic launch and the institution of ‘MAC’ pricing. We looked at a sample of generic switches from 2000 – 2009, and estimated the median time from first generic until first publication of a federal FUL was 13 months, though the time from generic launch to FUL fell steadily across this period (Exhibit 2). By 2009[16] FULs were appearing within about seven months of generic launch

Unlike AWP, AMP falls in parallel with acquisition costs, thus under an AMP-based benchmark, this ‘new generic’ period of very large dispensing margins is considerably reduced. AMP’s are published roughly 90 days in arrears[17]. For the first three months of a new generic, margins under AMP and AWP will be very similar – AMP will not be available for those first three months, thus AWP[18] is likely to remain the benchmark for the first three months following launch. AMP should appear in the fourth month following launch, and will be roughly equal to pharmacy’s acquisition costs in the first month after launch. From there, AMP tracks true acquisition costs with a +/- 90d lag. As such, unlike AWP, under AMP-based contracts the reimbursement paid by plan sponsors begins falling in or around the fourth month after a generic enters the market. Sellers still see more margin on newer generics because of the +/- 90d time lag inherent with AMP; on the assumption their acquisition costs are falling, sellers benefit from being paid a four month old (higher) price for a product they bought at the current (much lower) spot price. Nevertheless this ‘new generic’ margin is both smaller and shorter-lived under AMP than under AWP. Exhibit 3 compares pharmacy dispensing margin under AWP and AMP for a theoretical new generic with a $1.00 acquisition cost at launch. We assume AMP is unavailable for 3 months, and further assume that acquisition costs (and thus AMP) fall at the same rate as acquisition costs have fallen on average for the nine most significant recent generic launches[19]

How much sellers lose on new generics under AMP depends on how long we assume it would have taken for sellers to ‘MAC’ the new generic under the old AWP-based system. Crucially, PBMs don’t MAC their own mail order dispensing, so PBMs presumably lose the full gap[20] evident in Exhibit 3. Retailers would be subject to MAC pricing – presumably around month 7 – thus under AMP their losses on new generics are likely to be less than those experienced by PBMs[21]

  1. Center for Medicare and Medicaid Services
  2. Federal Register Volume 77, No. 22, Part II; February 2, 2012
  3. http://www.reginfo.gov/public/do/eAgendaViewRule?pubId=201210&RIN=0938-AQ41
  4. Nor is the final text of the rule. The rule published in February of 2012 is a proposed final rule, subject to comment. CMS has the option of incorporating comments or other changes into the final rule scheduled for publication in August 2013. Because replacing AWP with a publicly available benchmark that more closely tracks acquisition costs is required of CMS by language in the Affordable Care Act, we see no reason to believe the key elements of the proposed final rule, as they ultimately relate to generic dispensing margins under commercial contracts, will change materially
  5. AWP is simply a constant multiple (1.2x) of list price; list is also known as wholesale acquisition cost or ‘WAC’. Generic manufacturers post list (WAC) prices, then negotiate discounts to these list prices with wholesalers, retailers, and other large buyers. The magnitude of discounts is incredibly varied, and on average incredibly large. Thus AWP, which is 1.2x list, bears no reliable relationship to the actual transaction prices negotiated between generic manufacturers and the drug trades
  6. Government Accounting Office; see “Medicaid Outpatient Prescription Drugs”, December 15, 2010 (GAO-11-141R)
  7. Office of the Inspector General (for CMS); see “Analyzing Changes to Medicaid Federal Upper Limit Amounts”, October 2012 (OEI-03-11-00650)
  8. http://www.medicaid.gov/Medicaid-CHIP-Program-Information/By-Topics/Benefits/Prescription-Drugs/Downloads/reimbursementchart-2Q2013.pdf
  9. Even if a state chose to remain on AWP and attempted to stay within AMP-based FUL limits by simply reducing its re-imbursements as a percentage of AWP, the state still could not be sure it was within the AMP-based FUL limit. AMP bears a close and consistent relationship to true acquisition costs, whereas AWP is neither close nor consistent with respect to acquisition costs
  10. National Average Drug Acquisition Cost – a survey of retail pharmacies’ acquisition costs
  11. Drug retailers also are reliant on generic dispensing margins, though less so than PBMs; similarly drug wholesalers are reliant on generic dispensing margins, though less so than either drug retailers or PBMs
  12. SSR Health “PBM Pricing Post-AWP – An Estimate of Sustainable Earnings Power”, November 14, 2011
  13. AWP creates an information asymmetry that advantages the seller, but only in the context of generics, not brands. Accordingly, it makes sense for advantaged sellers to rely on this position of strength as much as is feasible
  14. Ibid 12
  15. Importantly, PBMs do not ‘MAC’ generics dispensed by their own mail order operations
  16. Federal FULs have not been updated or published since September 2009; we believe this is because of various legal uncertainties that existed in the period between CMS publication of new rules under the Deficit Reduction Act of 2005, and the Affordable Care Act based rules that are currently pending
  17. Manufacturers report AMP to CMS within 30d of the end of the preceding month, and CMS publishes those figures in the next month after they are received
  18. Wholesale acquisition cost (‘WAC’) may be used in the first 90d instead of AWP, but this has no practical relevance – WAC is simply AWP/1.2
  19. Specifically olanzapine; atorvastatin; escitalopram; quetiapine; clopidogrel; ziprasidone; irbesartan; fluvastatin; montelukast
  20. If we assume an equal number of claims are processed during each of the first 12 months, total dispensing margin for the new generic under AMP is 45 pct less than it would have been under AWP
  21. Assuming an equal number of prescriptions dispensed per month, MAC pricing at 7 months, and equivalency between MAC and 1.75 x AMP, retailers dispensing margin for the new generic is 18 pct less under AWP than under AMP
Print Friendly, PDF & Email